Summary
Tendencies towards smart grids change the circumstances for distribution network operators (DSOs) in the electricity sector. In order to fulfill their service obligations network operators can either rely on conventional approaches (based on copper) or choose more intelligent (so-called smart grid) solutions (based on the application of information and communication technologies). Under the German incentive regulation efficiency benchmarking plays a vital role resulting in individual efficiency targets for DSOs. In the context of regulatory benchmarking output parameters should adequately map service obligations of network operators. Against this background, this study analyses if new output parameters should be considered for the regulatory benchmarking in Germany.
One of the most crucial requirements regarding output parameters in the regulatory context is exogeneity. This means that outputs should not be influenced by decisions of network operators. A violation of this precondition might lead to distorted incentives for DSOs. The first three regulatory benchmarkings for German electricity DSOs show that the parameters network length and peak load do not fulfill the exogeneity precondition. Decisions are distorted in favor of conventional grid solutions.
Four possible solution have been identified in order to overcome this drawback. First, instead of using the actual peak load a certain quantile can be used as output for the regulatory benchmarking. This value could be linked to a pre-defined standard operation of the grid neglecting extraordinary peaks. Peak shaving measures would not affect output parameter values any more. Second, the connected load might be applied as output similar to considerations of differentiating between conditional and unconditional power demand. However, this approach requires an amendment of the existing conditions for network connections. Third, outputs with endogeneity problems can be related to pure exogenous parameters (like area supplied, connected customers etc.). These relations significantly reduce the endogeneity of the considered output. Last but not least, the econometric approach of instrumental variables (IV) can be applied. Instead of the actual values of the considered outputs, the values of the first step of an IV estimation would enter the regulatory benchmarking.
Further research is necessary in order to figure out, which of the four suggested ways is the most promising one. Finally, it should be underlined, that possible incentive distortions towards conventional grid measures should not be compensated by including explicit smart grid parameters (e.g. the number of installed intelligent transformer stations). Compensating one distortion by installing a second distortion with the opposite direction will not overcome the problem since you never know the magnitude of the distortions.
Diskussion Paper is available for download.
